Wind power mechanism prediction model training and life prediction method and device

Abstract

The invention provides a wind power mechanism prediction model training and life prediction method and device. The method comprises the steps of collecting operation data information of a wind power mechanism; performing associated factor extraction on the operation data information of the wind power mechanism to obtain associated feature information; inputting the associated feature information as local sample data into a deep convolutional neural network for training to obtain a regional wind power mechanism life prediction model; and performing global training to obtain a global wind power mechanism service life prediction model. Therefore, by collecting operation data information of a plurality of wind power mechanisms and analyzing and obtaining life correlation feature information of the wind power mechanisms, life prediction models of the wind power mechanisms in a plurality of regions are established, and finally, the life prediction models of the wind power mechanisms in the plurality of regions are globally trained to obtain a global life prediction model of the wind power mechanisms in the plurality of regions. Therefore, the problem of insufficient data coverage caused by limitation of historical data or a specific physical model can be avoided, and the accuracy of a prediction result is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of data analysis technology, and in particular to a method and device for training a wind power mechanism prediction model and predicting lifespan. Background Art

[0002] As the core equipment of wind power generation, the lifespan and operating conditions of wind turbines directly affect the efficiency and safety of wind power generation. In order to better understand the wear and aging of equipment, formulate reasonable maintenance plans, extend the service life of equipment, and improve economic benefits, it is very necessary to predict the lifespan of wind turbines. Existing methods for predicting the lifespan of wind turbines are mainly based on statistical analysis of historical data or simulation based on physical models. However, these methods only consider limited historical data or specific physical models when making predictions, and there is a problem of insufficient data coverage, which limits the accuracy of the prediction results. Therefore, the existing technology has a technical problem of insufficient accuracy in predicting the lifespan of wind turbines. Summary of the Invention

[0003] The present disclosure aims to solve one of the technical problems in the related art at least to a certain extent.

[0004] To this end, one purpose of the present disclosure is to propose a wind power structure life prediction model training method.

[0005] The second objective of the present disclosure is to provide a method for predicting the life of a wind power structure.

[0006] The third objective of the present disclosure is to provide a wind power structure life prediction model training device.

[0007] A fourth objective of the present disclosure is to provide a device for predicting the life of a wind power structure.

[0008] A fifth objective of the present disclosure is to provide an electronic device.

[0009] A sixth object of the present disclosure is to provide a non-transitory computer-readable storage medium.

[0010] A seventh object of the present disclosure is to provide a computer program product.

[0011] To achieve the above-mentioned purpose, the first embodiment of the present disclosure proposes a wind power mechanism life prediction model training method, including: collecting and obtaining multiple wind power mechanism operation data information in multiple regional environments; extracting correlation factors from the multiple wind power mechanism operation data information respectively to obtain multiple wind power mechanism life correlation feature information; inputting the multiple wind power mechanism life correlation feature information as local sample data into a deep convolutional neural network for training to obtain multiple regional wind power mechanism life prediction models; encrypting and transmitting the model parameters of the multiple regional wind power mechanism life prediction models to a federal central platform for global training to obtain a global wind power mechanism life prediction model.

[0012] According to one embodiment of the present disclosure, the extraction of correlation factors from the operating data information of the multiple wind power mechanisms to obtain multiple wind power mechanism life correlation feature information includes: obtaining wind power life influencing factors; constructing a wind power life feature decision tree based on the wind power life influencing factors; classifying the operating data information of the multiple wind power mechanisms based on the wind power life feature decision tree to obtain multiple wind power component classification feature sets; and fusing the multiple wind power component classification feature sets to obtain the multiple wind power mechanism life correlation feature information.

[0013] According to one embodiment of the present disclosure, constructing a wind power life characteristic decision tree includes: taking each factor of the wind power life influencing factors as a wind power classification feature in turn; performing information theory coding operations on the wind power classification features to obtain wind power classification feature information entropy; performing comparison based on the wind power classification feature information entropy to determine root node feature information; and constructing the wind power life characteristic decision tree based on the root node feature information.

[0014] According to one embodiment of the present disclosure, the method includes: obtaining feature information of decision tree leaf nodes based on the wind power life feature decision tree; performing influence analysis on the feature information of the decision tree leaf nodes to obtain root node influence information; if the root node influence information does not reach a preset influence threshold, pruning the wind power life feature decision tree.

[0015] According to one embodiment of the present disclosure, the life prediction models of wind power institutions in multiple regions are obtained, including: dividing the local sample data in proportion according to a preset sample ratio to determine a training set and a test set; adding data masks to the training set and the test set respectively; training the training set and the test set after adding the data masks through a deep convolutional neural network to obtain the life prediction models of wind power institutions in multiple regions.

[0016] According to one embodiment of the present disclosure, the method includes: verifying the life prediction models of wind power institutions in the multiple regions respectively to obtain multiple model prediction feedback data; performing loss analysis based on the multiple model prediction feedback data to obtain multiple predicted loss data; optimizing the life prediction models of wind power institutions in the multiple regions based on the multiple predicted loss data through a model optimization algorithm to generate multiple regional wind power institution life prediction optimization models.

[0017] To achieve the above-mentioned purpose, the second embodiment of the present disclosure proposes a wind power mechanism life prediction method, including: collecting target wind power mechanism operation data information of the target wind turbine, and obtaining a global wind power mechanism life prediction model, wherein the global wind power mechanism life prediction model is trained by the wind power mechanism life prediction model training method as described in the first embodiment; predicting the target wind power mechanism operation data information based on the global wind power mechanism life prediction model, outputting the wind power mechanism life prediction result, and maintaining and managing the wind power mechanism based on the wind power mechanism life prediction result.

[0018] According to one embodiment of the present disclosure, the maintenance and management of the wind power mechanism based on the wind power mechanism life prediction result includes: constructing a wind power mechanism operation and maintenance plan library; performing a maintenance level analysis on the wind power mechanism life prediction result to determine the wind power mechanism maintenance level information; matching the wind power mechanism maintenance level information with the wind power mechanism operation and maintenance plan library to obtain a target wind power mechanism operation and maintenance plan, and maintaining and managing the wind power mechanism through the target wind power mechanism operation and maintenance plan.

[0019] To achieve the above-mentioned purpose, the third aspect of the present disclosure proposes a wind power mechanism life prediction model training device, including: an acquisition module, used to collect and obtain multiple wind power mechanism operation data information in multiple regional environments; an extraction module, used to extract correlation factors from the multiple wind power mechanism operation data information respectively, and obtain multiple wind power mechanism life correlation feature information; a first training module, used to input the multiple wind power mechanism life correlation feature information as local sample data into a deep convolutional neural network for training, and obtain multiple regional wind power mechanism life prediction models; a second training module, used to encrypt and transmit the model parameters of the multiple regional wind power mechanism life prediction models to a federal central platform for global training, so as to obtain a global wind power mechanism life prediction model.

[0020] To achieve the above-mentioned purpose, the fourth embodiment of the present disclosure proposes a wind power mechanism life prediction device, including: an acquisition module, used to collect target wind power mechanism operation data information of a target wind turbine, and obtain a global wind power mechanism life prediction model, wherein the global wind power mechanism life prediction model is obtained by training through the wind power mechanism life prediction model training method as described in the first embodiment; a prediction module, used to predict the target wind power mechanism operation data information based on the global wind power mechanism life prediction model, output the wind power mechanism life prediction result, and maintain and manage the wind power mechanism based on the wind power mechanism life prediction result.

[0021] To achieve the above-mentioned purpose, the fifth aspect embodiment of the present disclosure proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to implement the wind power mechanism life prediction model training method as described in the first aspect embodiment of the present disclosure.

[0022] To achieve the above-mentioned purpose, the sixth embodiment of the present disclosure proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to implement the wind power mechanism life prediction model training method as described in the first embodiment of the present disclosure.

[0023] To achieve the above-mentioned purpose, the seventh embodiment of the present disclosure proposes a computer program product, including a computer program, which, when executed by a processor, is used to implement the wind power structure life prediction model training method as described in the first embodiment of the present disclosure.

[0024] Therefore, by collecting the operating data information of multiple wind power institutions and analyzing and obtaining the characteristic information related to the life of wind power institutions, a life prediction model of wind power institutions in multiple regions is established. Finally, by globally training the life prediction models of wind power institutions in multiple regions, a global wind power institution life prediction model is obtained. This can avoid the problem of insufficient data coverage due to the limitations of historical data or specific physical models and improve the accuracy of the prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic diagram of a wind power unit life prediction model training method according to an embodiment of the present disclosure;

[0026] Figure 2 is a schematic diagram of another wind power structure life prediction model training method according to an embodiment of the present disclosure;

[0027] Figure 3is a schematic diagram of another wind power structure life prediction model training method according to an embodiment of the present disclosure;

[0028] Figure 4 is a schematic diagram of a wind power structure life prediction method according to one embodiment of the present disclosure;

[0029] Figure 5 This is a schematic diagram of a wind power mechanism life prediction model training device according to one embodiment of the present disclosure;

[0030] Figure 6 This is a schematic diagram of a wind power mechanism life prediction device according to one embodiment of the present disclosure;

[0031] Figure 7 is a schematic diagram of an electronic device according to one embodiment of the present disclosure. DETAILED DESCRIPTION

[0032] The following describes in detail embodiments of the present disclosure, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, and should not be construed as limiting the present disclosure.

[0033] The acquisition, storage, use, and processing of data in the technical solution disclosed herein are in compliance with the relevant provisions of relevant laws and regulations.

[0034] Figure 1 FIG. 1 is a schematic diagram of a wind power unit life prediction model training method according to an embodiment of the present disclosure. Figure 1 As shown, the wind power structure life prediction model training method includes the following steps:

[0035] S101, collecting and obtaining operating data information of multiple wind power units in multiple regional environments.

[0036] The wind power mechanism life prediction model training method of the embodiment of the present application can be applied to the scenario of wind power mechanism equipment maintenance. The executor of the wind power mechanism life prediction model training of the embodiment of the present application can be the wind power mechanism life prediction model training device of the embodiment of the present application, and the wind power mechanism life prediction model training device can be set on an electronic device.

[0037] In the embodiment of the present disclosure, there may be various methods for collecting and acquiring the operating data information of multiple wind power units in multiple regional environments, which are not limited herein.

[0038] In one possible implementation, operating data of wind turbines in multiple regions can be collected through sensors, monitoring systems, and other means, including environmental factors such as wind speed, wind direction, temperature, humidity, and air pressure, as well as data such as the operating status and fault records of the wind turbines. A wind turbine refers to an important component of a wind turbine generator set, including a rotor, gear unit, generator, speed regulating device, brake mechanism, and other parts. The wind turbine is the core component of a wind turbine generator set, and its operating status and lifespan directly affect the efficiency and reliability of the entire wind power generation system. Multiple regions refer to the different geographical locations of wind turbines, which are distributed in different regions. By collecting operating data of wind turbines in multiple regions, we can fully understand the operating conditions and life-influencing factors of wind turbines in different regions, thereby accurately predicting the lifespan of wind turbines.

[0039] In the disclosed embodiments, real-time monitoring and collection of wind turbine operating data can be achieved by installing a sensor group on a target wind turbine. The target wind turbine refers to the wind turbine whose operating data is monitored and collected in real time by the sensor group. This can be a single wind turbine or multiple wind turbines in a wind farm. The sensor group includes various types of sensors, such as temperature sensors, humidity sensors, and wind speed sensors, used to capture and measure data from various aspects of the wind turbine.

[0040] It should be noted that there may be many factors that affect the life of wind turbines, which are not limited here. For example, they may include operating environment, mechanism type, component quality, wind turbine material, etc.

[0041] In another possible implementation, the operation data information of the corresponding wind power unit may be determined by obtaining the operation log of the wind power station or wind power equipment.

[0042] S102 , extracting correlation factors from the operating data information of multiple wind power units respectively to obtain life-span correlation feature information of multiple wind power units.

[0043] In the embodiment of the present disclosure, the collected operating data information of multiple wind turbines can be processed and analyzed, and characteristic information related to the life of the wind turbine can be extracted through correlation analysis and other indicators. For example, environmental factors such as wind speed, wind direction, and temperature are related to the life of the wind turbine. Excessive wind speed will cause increased wear of the wind rotor blades, thereby shortening their life. Temperature changes will also affect the operating efficiency of the wind turbine and the life of its components. These factors are related factors related to the life of the wind turbine. In addition, the operating status and fault records of the wind turbine will also affect the life of the wind turbine. If the wind turbine often fails or the maintenance records show that certain components are frequently replaced, then this information indicates that the life of the wind turbine is short. Therefore, these factor data can also be used as characteristic information. By extracting these related factors, multiple wind turbine life-related characteristic information can be constructed.

[0044] It should be noted that there are many methods for extracting correlation factors, which are not limited here and can be specifically limited according to actual design needs.

[0045] In one possible implementation, life-related characteristic information of multiple wind power units may be obtained from operation data of multiple wind power units through expert evaluation.

[0046] In another possible implementation, a correlation factor extraction model can be used to process operational data from multiple wind turbines to obtain lifespan-related feature information for the multiple wind turbines. This correlation factor extraction model is pre-trained and stored in the storage space of an electronic device for easy access when needed.

[0047] S103: Inputting the life-related characteristic information of multiple wind power units as local sample data into a deep convolutional neural network for training to obtain life prediction models of wind power units in multiple regions.

[0048] The extracted features associated with the lifespans of multiple wind turbines are used as local sample data and fed into a deep convolutional neural network for training. A deep convolutional neural network is a powerful deep learning model that uses these features as input and, through training, learns the mapping from these features to wind turbine lifespans. This results in wind turbine lifespan prediction models for multiple regions. These models are applicable to wind turbines in different regions and can predict the lifespans of corresponding wind turbines based on the new features input. By training these models for wind turbine lifespan predictions in multiple regions, more accurate and reliable wind turbine lifespan predictions can be achieved.

[0049] S104: Encrypt and transmit the model parameters of the wind power plant life prediction models of multiple regions to the federal central platform for global training to obtain a global wind power plant life prediction model.

[0050] It should be noted that because different regions have different wind turbine lifespan-related feature information or different levels of importance, the regional wind turbine lifespan prediction models for different regions may be different. To achieve the goal of predicting the lifespan of wind turbines in all regions, it is necessary to globally train the wind turbine lifespan prediction models for multiple regions to obtain a global wind turbine lifespan prediction model.

[0051] In the disclosed embodiment, the model parameters of the wind turbine life prediction model for each region can be encrypted and transmitted to the federal central platform for global training. The federal central platform uses extensive data and knowledge to construct a more accurate and reliable global wind turbine life prediction model. This model takes into account the data differences and characteristics between different regions, can discover and utilize data associations and patterns between different regions, and by integrating wind turbine operation data and experience from different regions, better captures the relationship between wind turbine life and various factors, and achieves comprehensive and objective life prediction. This model is applicable to wind turbines in all regions. Through global training, it avoids the problems of overfitting and insufficient generalization ability that may exist in local models, thereby improving the adaptability and accuracy of the model.

[0052] In the embodiment of the present disclosure, first, the operating data information of multiple wind power units in multiple regional environments is collected and obtained, and then the correlation factors of the operating data information of the multiple wind power units are extracted respectively to obtain the characteristic information related to the life of multiple wind power units. Then, the characteristic information related to the life of multiple wind power units is input into the deep convolutional neural network as local sample data for training to obtain the life prediction models of multiple regional wind power units. Finally, the model parameters of the life prediction models of the multiple regional wind power units are encrypted and transmitted to the federal central platform for global training to obtain the global wind power unit life prediction model. Thus, by collecting the operating data information of multiple wind power units and analyzing and obtaining the characteristic information related to the life of wind power units, the life prediction models of multiple regional wind power units are established. Finally, by globally training the life prediction models of multiple regional wind power units, the global wind power unit life prediction model is obtained. This can avoid the problem of insufficient data coverage due to the limitations of historical data or specific physical models, and improve the accuracy of the prediction results.

[0053] In the above embodiment, the operating data information of multiple wind power units are respectively subjected to correlation factor extraction to obtain the life-related characteristic information of multiple wind power units. Figure 2 Explaining further, the method includes:

[0054] S201, obtaining factors affecting wind turbine lifespan.

[0055] It should be noted that there are many factors that affect the lifespan of wind turbines, and no limitation is given here. For example, they may include the operating environment, mechanism type, component quality, wind turbine material, etc. These factors will have a direct or indirect impact on the lifespan of wind turbines. Among them, the operating environment of the wind turbine has a direct impact on its lifespan. For example, extreme weather conditions (such as storms, ice and snow, etc.) may damage the wind turbine, reducing its operating efficiency and lifespan. Different types of wind turbines have differences in design and structure, and their lifespans will also be affected differently. For example, different types of wind turbines (such as horizontal axis wind turbines and vertical axis wind turbines) have differences in operating efficiency and lifespan. The quality of the components in the wind turbine has a significant impact on its lifespan. If there are quality problems with the components, such as manufacturing defects and material aging, it may cause the wind turbine to malfunction during operation, reducing its lifespan. The material of the wind turbine also has a significant impact on its lifespan. For example, materials such as steel and aluminum alloy may be affected by corrosion and wear during long-term operation, thereby reducing its lifespan.

[0056] S202: Construct a wind power life characteristic decision tree based on factors affecting wind power life.

[0057] It should be noted that a decision tree is a supervised learning algorithm based on machine learning, commonly used for classification and regression analysis tasks. It represents possible decisions, outcomes, and their probabilities in a tree-like model. By learning the relationships between features in a dataset and a target variable, it constructs a tree-like structure consisting of internal nodes (test nodes), branches (decision rules), and leaf nodes (predicted results). In a decision tree, each internal node represents a test of a feature or attribute, each branch represents a possible value for that feature, and each leaf node represents a category or a continuous numerical result. The process of constructing a decision tree typically involves partitioning the dataset, selecting optimal features, and generating child nodes. Metrics such as information entropy, Gini impurity, and chi-square tests are used to evaluate and select the optimal split point. Decision trees are easy to understand and interpret, can handle both discrete and continuous features, and can visualize the decision-making process. Therefore, they are widely used in fields such as financial risk control, marketing strategy development, and medical diagnosis. However, decision trees also have overfitting problems, which can be optimized through pruning and other methods, as well as ensemble learning methods such as Random Forest to further improve model performance and generalization ability.

[0058] In the disclosed embodiment, a wind turbine lifespan feature decision tree can be constructed based on the acquired factors influencing wind turbine lifespan. The data can be divided into different subsets based on the features using algorithms such as the Classification and Regression Trees (CART) algorithm and the Iterative Binary Tree 3 (ID3) algorithm, and a decision tree can be recursively constructed. For example, the decision tree may divide wind turbines into two categories based on their environment (coastal or inland). Within each category, further classification is performed based on the type of wind turbine (horizontal axis or vertical axis), ultimately resulting in a decision tree encompassing wind turbines of different categories.

[0059] In one possible implementation method, each factor influencing the life of wind power can be used as a wind power classification feature in turn, and then the wind power classification features are subjected to information theory coding operations to obtain the wind power classification feature information entropy. Then, based on the wind power classification feature information entropy, a comparison is performed to determine the root node feature information. Finally, according to the root node feature information, a wind power life feature decision tree is constructed.

[0060] Specifically, for each wind power classification feature, the encoding function in information theory can be used to convert it into the corresponding amount of information. The uncertainty or information amount of each feature is calculated by comparing the distribution of features in different categories to evaluate its importance. The information entropy formula is used to calculate the information entropy of the wind power classification feature, which is used to evaluate the importance of each feature.

[0061] After calculating the information entropy of each wind power classification feature, the information entropies of each wind power classification feature are compared. Since lower information entropy indicates higher certainty, that is, the feature has a stronger indicative effect on the classification result, the feature with the smallest information entropy is selected as the root node feature. The classification based on this feature can minimize the uncertainty of subsequent classification and ensure that the decision tree can more accurately predict the life of the wind power unit in the subsequent branch generation process.

[0062] In another possible implementation, it is also possible to start from the root node of the wind turbine life characteristic decision tree, traverse downward along the path of the decision tree until reaching a leaf node, and obtain the influencing factors corresponding to the leaf node, that is, the feature information of the leaf node of the decision tree.

[0063] Calculate the number and proportion of samples for each leaf node. Based on the number and proportion of samples for each leaf node, calculate the contribution of each leaf node to the root node. This is the number or proportion of samples for the leaf node multiplied by the importance of the influencing factor represented by the node. Combined with the contributions of all leaf nodes, the total influence of the root node is obtained.

[0064] An influence threshold is set. The influence threshold is a threshold used to determine whether a leaf node is important. It is set based on experience, data distribution, business needs and other factors. Starting from the root node of the decision tree, an impact assessment is performed on each branch. If the influence of a branch is too small and does not reach the preset influence threshold, it indicates that the leaf node corresponding to the branch is not important and it is pruned. Repeat the above steps until the influence of all branches reaches the preset influence threshold or no more branches can be pruned. This preferred embodiment optimizes the structure of the decision tree by pruning unimportant leaf nodes, reduces the uncertainty of classification, and reduces the computational complexity of the decision tree, thereby achieving the technical effect of optimizing the structure and performance of the decision tree.

[0065] S203 , classifying the operating data information of multiple wind power units based on the wind power life feature decision tree to obtain multiple wind power component classification feature sets.

[0066] In the disclosed embodiments, the acquired operational data for multiple wind turbines can be input into a constructed wind turbine lifespan feature decision tree. Based on the decision tree's classification rules, the operational data for each wind turbine is classified to obtain multiple wind turbine component classification feature sets. These sets reflect the characteristic performance of each component in different categories. For example, the operational data for a horizontal-axis wind turbine located on the coast might be classified into a combination of "coastal environment" and "horizontal-axis wind turbine."

[0067] It should be noted that the classification rules are set in advance and can be changed according to actual design needs. No restrictions are imposed here.

[0068] S204: Merge the classification feature sets of multiple wind power components to obtain life-related feature information of multiple wind power units.

[0069] In the embodiment of the present disclosure, there are many ways to fuse the classification feature sets of multiple wind turbine components, which are not limited here. For example, the average value, maximum value, minimum value, etc. of each feature set can be calculated, and through fusion, the life-related feature information of multiple wind turbine mechanisms can be obtained, that is, a comprehensive feature vector is obtained. This vector contains the life-related feature information of all wind turbine mechanisms and comprehensively reflects the life and performance of the wind turbine mechanisms. For example, suppose there are two classification feature sets of wind turbine components: one represents the effect of temperature on life, and the other represents the effect of humidity on life. By fusing these two sets, a comprehensive feature vector is obtained. This vector takes into account the effect of temperature and humidity on the life of the wind turbine mechanism. This preferred embodiment obtains a comprehensive feature vector by fusing multiple classification feature sets. This vector contains the life-related feature information of all wind turbine mechanisms, thereby achieving the technical effect of comprehensively obtaining the life-related feature information of wind turbine mechanisms.

[0070] Based on the root node features, the decision tree branches are recursively constructed. For each branch, a subset of features is selected for further classification. When selecting subset features, information entropy is still used as the evaluation criterion, and the feature that can minimize the uncertainty of the subset classification is selected as the feature of the next node. This process is repeated until a certain termination condition is reached (such as reaching a preset depth or meeting a certain threshold), completing the construction of the wind turbine life feature decision tree. This preferred embodiment selects the feature with the smallest information entropy as the root node, allowing the decision tree to minimize the uncertainty of subsequent classification, thereby accurately predicting the life of the wind turbine structure, reducing the risk of misjudgment, and achieving the technical effect of improving classification accuracy.

[0071] In the disclosed embodiment, factors influencing wind turbine lifespan are first obtained. These factors include operating environment, mechanism type, component quality, and wind turbine material. A wind turbine lifespan feature decision tree is then constructed based on these factors. Based on the wind turbine lifespan feature decision tree, multiple wind turbine mechanism operating data are classified to obtain multiple wind turbine component classification feature sets. Finally, these multiple wind turbine component classification feature sets are fused to obtain multiple wind turbine mechanism lifespan-related feature information. Thus, by establishing a wind turbine lifespan feature decision tree, classification uncertainty can be reduced, the wind turbine mechanism lifespan can be accurately predicted, the risk of misjudgment can be reduced, and classification accuracy can be improved.

[0072] In the above embodiment, the life prediction models of wind power units in multiple regions are obtained, and the prediction models can also be obtained by Figure 3 Explaining further, the method includes:

[0073] S301: Divide the local sample data into proportions according to a preset sample ratio to determine a training set and a test set.

[0074] It should be noted that the preset sample ratio can be changed according to actual design needs and is not limited here. For example, the preset sample ratio can be 10:1 for the number of training sets: the number of test sets.

[0075] S302, adding data masks to the training set and the test set respectively.

[0076] In the disclosed embodiments, a data masking strategy can be used to add data masks to the training and test sets, respectively. Data masking is a technique used to protect data privacy and security. After adding data masking, certain parts of the original data are obscured or replaced to prevent unauthorized access or use. This is used to hide sensitive information, such as geographic location and device model, to prevent the disclosure of sensitive information. For example, different random number generators are used to generate the masks to ensure that the data distribution between the training and test sets is similar.

[0077] It should be noted that the data masking strategy is designed in advance and can be changed according to actual design requirements, and is not limited here.

[0078] S303: Training the training set and the test set after adding the data mask through a deep convolutional neural network to obtain a life prediction model for wind power units in multiple regions.

[0079] In the disclosed embodiments, a deep convolutional neural network can be used to train a model using a data-masked training set, and to evaluate the model's performance using a data-masked test set. Data is fed into the network, and the network's weights are updated using a backpropagation algorithm. This process continues until the model's performance reaches a satisfactory level. The model's performance is evaluated by calculating metrics such as accuracy, recall, and F1 score.

[0080] It can be understood that model training is an iterative process, which is carried out by continuously adjusting the network parameters of the model until the overall loss function value of the model is less than the preset value, or the overall loss function value of the model no longer changes or changes slowly, the model converges, and a trained model is obtained.

[0081] In the disclosed embodiments, a deep convolutional neural network can be used to train a model using a data-masked training set, and to evaluate the model's performance using a data-masked test set. Data is fed into the network, and the network's weights are updated using a backpropagation algorithm. This process continues until the model's performance reaches a satisfactory level. The model's performance is evaluated by calculating metrics such as accuracy, recall, and F1 score.

[0082] After achieving satisfactory performance on the test set, the model generated lifespan prediction models for wind turbines in multiple regions. These models are used to predict the lifespans of wind turbines in different regions. New wind turbine data is fed into the model, which outputs a predicted lifespan based on the characteristics and patterns learned from the historical data. This preferred implementation utilizes a data masking strategy to hide sensitive information, preventing its leakage and achieving the technical effect of protecting data privacy and security.

[0083] After the training is completed, the life prediction models of wind power institutions in multiple regions can be verified separately to obtain multiple model prediction feedback data. Then, loss analysis can be performed based on the multiple model prediction feedback data to obtain multiple predicted loss data. Finally, the life prediction models of wind power institutions in multiple regions can be optimized based on the multiple predicted loss data through the model optimization algorithm to generate multiple regional wind power institution life prediction optimization models.

[0084] The trained wind power unit life prediction models for multiple regions are verified separately. The models are verified using data from test sets or actual scenarios, the performance of the models on unseen data is evaluated, and prediction feedback data for each model is collected, including the model's prediction results, actual results, and model errors.

[0085] Based on the prediction feedback data from multiple models, we calculate the loss function (mean square error, root mean square error, etc.). The loss function is a metric used to measure the model's prediction error; smaller values ​​indicate smaller prediction errors. By calculating the loss function, we obtain multiple prediction loss data. This data is used to evaluate the model's performance and prediction error. By comparing the loss function values ​​of different models, we can determine which models perform better in terms of prediction error.

[0086] After calculating the loss function, the gradient of the loss function is calculated. The gradient is the partial derivative of the loss function with respect to the model parameters, which is used to guide the update direction of the model parameters. According to the gradient of the loss function, an optimization algorithm (such as gradient descent method, stochastic gradient descent method, etc.) is used to update the parameters of the model, including adjusting the parameters of the model, improving the architecture of the model, adding new features, etc., to minimize the loss function and improve the performance of the model. After each update of the model parameters, the value and gradient of the loss function are recalculated, and the optimization process is continued until a certain convergence condition is reached or a preset number of iterations is reached. Through the optimization process of the model optimization algorithm, multiple regional wind power structure life prediction optimization models are generated. These optimized models have better performance and lower prediction errors. This preferred embodiment finds the optimal model parameters through the model optimization algorithm, improves the prediction accuracy of the model, and thus achieves the technical effect of improving the accuracy of wind power structure life prediction.

[0087] In this disclosed embodiment, local sample data is first divided according to a preset sample ratio to determine training and test sets. Data masks are then added to the training and test sets, respectively. Finally, a deep convolutional neural network is trained on the masked training and test sets to generate lifespan prediction models for wind turbines in multiple regions. By masking the training and test sets, sensitive information can be prevented from being leaked, enhancing data confidentiality.

[0088] Figure 4 FIG. 1 is a schematic diagram of a method for predicting the life of a wind power unit according to an embodiment of the present disclosure. Figure 4 As shown, the wind power mechanism life prediction method includes the following steps:

[0089] S401 , collecting target wind power mechanism operating data information of a target wind turbine, and obtaining a global wind power mechanism life prediction model.

[0090] It should be noted that the global wind power mechanism life prediction model in the embodiment of the present disclosure is as follows: Figure 1-Figure 3 The wind power structure life prediction model training method shown is trained.

[0091] S402 , predicting the target wind power unit operating data information based on the global wind power unit life prediction model, outputting the wind power unit life prediction result, and performing maintenance and management on the wind power unit based on the wind power unit life prediction result.

[0092] In the embodiment of the present disclosure, the target wind turbine operating data information of the target wind turbine is first collected, and a global wind turbine life prediction model is obtained. Then, the target wind turbine operating data information is predicted based on the global wind turbine life prediction model, and the wind turbine life prediction result is output. The wind turbine maintenance management is performed based on the wind turbine life prediction result. Figure 1-Figure 3 The global wind power mechanism life prediction model obtained by training the wind power mechanism life prediction model training method shown in the figure predicts the life of the target wind power mechanism operation data information of the target wind turbine, which can improve the prediction effect and accuracy.

[0093] In the disclosed embodiment, the wind power mechanism is maintained and managed based on the wind power mechanism life prediction result, and a wind power mechanism operation and maintenance plan library can also be constructed. Then, a maintenance level analysis is performed on the wind power mechanism life prediction result to determine the wind power mechanism maintenance level information. Finally, the wind power mechanism maintenance level information is matched with the wind power mechanism operation and maintenance plan library to obtain a target wind power mechanism operation and maintenance plan, and the wind power mechanism is maintained and managed through the target wind power mechanism operation and maintenance plan.

[0094] Specifically, the operation and maintenance data and experience of wind power organizations can be collected, including various maintenance methods, maintenance tools, maintenance time, maintenance effects, etc. Based on the collected data and experience, various operation and maintenance plans can be formulated, including preventive maintenance, predictive maintenance, troubleshooting, etc. These operation and maintenance plans can be organized into a library to form a wind power organization operation and maintenance plan library for subsequent matching and selection.

[0095] Then, based on the life prediction results of the wind power mechanism, the remaining service life and reliability of each equipment are evaluated. Based on the remaining service life and reliability, the maintenance level of the wind power mechanism is analyzed and the wind power mechanism is divided into different maintenance levels, such as high risk, medium risk and low risk levels.

[0096] When matching plans, matching algorithms (fuzzy matching, rule matching, etc.) can be used to match the maintenance level information of wind power organizations with the operation and maintenance plan library. For example, for wind power organizations with high risk levels, preventive maintenance and predictive maintenance plans may be matched to reduce the probability of failures; for wind power organizations with medium risk levels, preventive maintenance, predictive maintenance plans may be matched, or troubleshooting plans may be selected based on actual conditions; for wind power organizations with low risk levels, troubleshooting plans may be matched.

[0097] Based on the matching results, a target O&M plan is determined for each wind turbine. Maintenance and management of the wind turbine is then implemented using this target O&M plan. This includes developing specific maintenance plans and operational guidelines based on the target O&M plan using project management tools, flowcharts, and other methods. Regular maintenance and inspections are performed on the wind turbines according to these plans and guidelines, and data is promptly recorded and maintained during the maintenance process using spreadsheets and databases. This preferred implementation achieves the technical effect of improving the standardization and regularization of wind turbine O&M management through the construction of an O&M plan library, the analysis and matching of maintenance levels, and the development of specific maintenance plans and operational guidelines.

[0098] Corresponding to the wind power mechanism life prediction model training method provided in the above-mentioned embodiments, an embodiment of the present disclosure also provides a wind power mechanism life prediction model training device. Since the wind power mechanism life prediction model training device provided in the embodiment of the present disclosure corresponds to the wind power mechanism life prediction model training method provided in the above-mentioned embodiments, the implementation method of the above-mentioned wind power mechanism life prediction model training method is also applicable to the wind power mechanism life prediction model training device provided in the embodiment of the present disclosure, and will not be described in detail in the following embodiments.

[0099] Figure 5 FIG. 1 is a schematic diagram of a wind power mechanism life prediction model training device according to an embodiment of the present disclosure. Figure 5As shown, the wind power structure life prediction model training device 500 includes: a collection module 510, an extraction module 520, a first training module 530 and a second training module 540.

[0100] The acquisition module 510 is used to acquire operating data information of multiple wind power units in multiple regional environments.

[0101] The extraction module 520 is used to extract correlation factors from the operating data information of multiple wind power units respectively, and obtain the life-span correlation feature information of multiple wind power units.

[0102] The first training module 530 is used to input the life-related feature information of multiple wind power units as local sample data into a deep convolutional neural network for training, so as to obtain life prediction models of wind power units in multiple regions.

[0103] The second training module 540 is used to encrypt and transmit the model parameters of the wind power plant life prediction models of multiple regions to the federal central platform for global training to obtain a global wind power plant life prediction model.

[0104] In one embodiment of the present disclosure, the extraction module 520 is further used to: obtain factors affecting wind power life; construct a wind power life feature decision tree based on the factors affecting wind power life; classify the operating data information of multiple wind power units based on the wind power life feature decision tree to obtain multiple wind power component classification feature sets; and fuse the multiple wind power component classification feature sets respectively to obtain multiple wind power unit life-related feature information.

[0105] In one embodiment of the present disclosure, the extraction module 520 is further used to: sequentially use each factor among the factors affecting wind power life as a wind power classification feature; perform information theory coding operations on the wind power classification features to obtain wind power classification feature information entropy; perform comparison based on the wind power classification feature information entropy to determine the root node feature information; and construct a wind power life feature decision tree based on the root node feature information.

[0106] In one embodiment of the present disclosure, the extraction module 520 is further used to: obtain feature information of decision tree leaf nodes based on the wind power life feature decision tree; perform influence analysis on the feature information of the decision tree leaf nodes to obtain root node influence information; and prune the wind power life feature decision tree if the root node influence information does not reach a preset influence threshold.

[0107] In one embodiment of the present disclosure, the first training module 530 is further used to: divide the local sample data proportionally according to a preset sample ratio to determine a training set and a test set; add data masks to the training set and the test set respectively; and train the training set and the test set with the added data masks through a deep convolutional neural network to obtain life prediction models for wind power units in multiple regions.

[0108] In one embodiment of the present disclosure, the first training module 530 is further used to: verify the life prediction models of wind power institutions in multiple regions respectively to obtain multiple model prediction feedback data; perform loss analysis based on the multiple model prediction feedback data to obtain multiple predicted loss data; optimize the life prediction models of wind power institutions in multiple regions based on the multiple predicted loss data through a model optimization algorithm to generate multiple regional wind power institution life prediction optimization models.

[0109] In one embodiment of the present disclosure, the first training module 530 is further used to: verify the life prediction models of wind power institutions in multiple regions respectively to obtain multiple model prediction feedback data; perform loss analysis based on the multiple model prediction feedback data to obtain multiple predicted loss data; optimize the life prediction models of wind power institutions in multiple regions based on the multiple predicted loss data through a model optimization algorithm to generate multiple regional wind power institution life prediction optimization models.

[0110] Therefore, by collecting the operating data information of multiple wind power institutions and analyzing and obtaining the characteristic information related to the life of wind power institutions, a life prediction model of wind power institutions in multiple regions is established. Finally, by globally training the life prediction models of wind power institutions in multiple regions, a global wind power institution life prediction model is obtained. This can avoid the problem of insufficient data coverage due to the limitations of historical data or specific physical models and improve the accuracy of the prediction results.

[0111] Corresponding to the wind power mechanism life prediction methods provided in the above-mentioned embodiments, an embodiment of the present disclosure also provides a wind power mechanism life prediction device. Since the wind power mechanism life prediction device provided in the embodiment of the present disclosure corresponds to the wind power mechanism life prediction methods provided in the above-mentioned embodiments, the implementation methods of the above-mentioned wind power mechanism life prediction methods are also applicable to the wind power mechanism life prediction device provided in the embodiment of the present disclosure, and will not be described in detail in the following embodiments.

[0112] Figure 6 Schematic diagram of a wind power mechanism life prediction device according to one embodiment of the present disclosure. Figure 6 As shown, the wind power mechanism life prediction device 600 includes: an acquisition module 610 and a prediction module 620.

[0113] The acquisition module 610 is used to collect target wind turbine operating data information and obtain a global wind turbine life prediction model. The global wind turbine life prediction model is obtained by Figure 1-Figure 3 The wind power structure life prediction model training method shown is trained.

[0114] The prediction module 620 is used to predict the target wind power unit operating data information based on the global wind power unit life prediction model, output the wind power unit life prediction result, and perform maintenance and management on the wind power unit based on the wind power unit life prediction result.

[0115] In one embodiment of the present disclosure, the prediction module 620 is further used to: construct a wind power mechanism operation and maintenance solution library; perform maintenance level analysis on the wind power mechanism life prediction results to determine the wind power mechanism maintenance level information; match the wind power mechanism maintenance level information with the wind power mechanism operation and maintenance solution library to obtain a target wind power mechanism operation and maintenance solution, and perform maintenance management on the wind power mechanism using the target wind power mechanism operation and maintenance solution. Figure 1-Figure 3 The global wind power mechanism life prediction model obtained by training the wind power mechanism life prediction model training method shown in the figure predicts the life of the target wind power mechanism operation data information of the target wind turbine, which can improve the prediction effect and accuracy.

[0116] In order to implement the above embodiment, the present disclosure further provides an electronic device 700. Figure 7 is a schematic diagram of an electronic device according to an embodiment of the present disclosure, such as Figure 7 As shown, the electronic device 700 includes: a processor 701 and a memory 702 in communication with the processor, the memory 702 stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor 701 to implement the present disclosure. Figure 1-Figure 3 The wind power mechanism life prediction model training method of the embodiment, or Figure 4 A method for predicting the life of a wind power structure according to an embodiment.

[0117] In order to implement the above embodiment, the present disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to implement the above embodiment. Figure 1-Figure 3 The wind power mechanism life prediction model training method of the embodiment, or Figure 4 A method for predicting the life of a wind power structure according to an embodiment.

[0118] In order to implement the above embodiments, the present disclosure also provides a computer program product, including a computer program, which implements the above embodiments when executed by a processor. Figure 1-Figure 3 The wind power mechanism life prediction model training method of the embodiment, or Figure 4 A method for predicting the life of a wind power structure according to an embodiment.

[0119] It is important to note that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold beyond these legitimate uses. Furthermore, such collection / sharing should be conducted only after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes the relevant user information before using the feature. Furthermore, any necessary steps must be taken to safeguard and secure access to such personal information and ensure that others with access to personal information comply with its privacy policy and procedures.

[0120] This application contemplates providing implementations that allow users to selectively block the use or access of personal information data. Specifically, this disclosure contemplates providing hardware and / or software to prevent or block access to such personal information data. Risks can be minimized by limiting data collection and deleting data once it is no longer needed. Furthermore, where applicable, such personal information can be de-identified to protect user privacy.

[0121] In the descriptions of the foregoing embodiments, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are mutually inconsistent.

[0122] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. Throughout the description of this application, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0123] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0124] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing it in another suitable manner if necessary, and then storing it in a computer memory.

[0125] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0126] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0127] In addition, the functional units in the various embodiments of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into a module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0128] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A wind power unit life prediction model training method, characterized in that: The method comprises: Collect and obtain operating data information of multiple wind power organizations in multiple regional environments; Extracting correlation factors from the operating data information of the plurality of wind power units respectively to obtain life-span correlation feature information of the plurality of wind power units; Inputting the life-related characteristic information of the plurality of wind power units as local sample data into a deep convolutional neural network for training, thereby obtaining life prediction models of wind power units in multiple regions; The model parameters of the wind power plant life prediction models in the multiple regions are encrypted and transmitted to the federal central platform for global training to obtain a global wind power plant life prediction model.

2. The method according to claim 1, characterized in that The extracting of correlation factors from the operating data of the plurality of wind power units to obtain the life-span correlation characteristic information of the plurality of wind power units includes: Obtain factors affecting wind turbine lifespan; Constructing a wind power life characteristic decision tree based on the factors affecting the wind power life; Classifying the plurality of wind power mechanism operation data information based on the wind power life characteristic decision tree to obtain a plurality of wind power component classification feature sets; The plurality of wind power component classification feature sets are respectively fused to obtain the plurality of wind power mechanism life-span related feature information.

3. The method according to claim 2, characterized in that The construction of the wind power life characteristic decision tree includes: Taking each factor influencing the life of wind power as a wind power classification feature in turn; Performing information theory coding operations on the wind power classification features to obtain wind power classification feature information entropy; Performing a comparison based on the wind power classification feature information entropy to determine the root node feature information; The wind power life characteristic decision tree is constructed according to the root node characteristic information.

4. The method according to claim 3, characterized in that The method comprises: Obtaining feature information of tree nodes of a decision tree based on the wind power life feature decision tree; Performing influence analysis on the feature information of the decision tree leaves to obtain root node influence information; If the root node influence information does not reach a preset influence threshold, the wind power life characteristic decision tree is pruned.

5. The method according to claim 1, wherein The life prediction models of wind power plants in multiple regions are obtained, including: Divide the local sample data into a proportion according to a preset sample ratio to determine a training set and a test set; Adding data masks to the training set and the test set respectively; The training set and the test set after adding the data mask are trained through a deep convolutional neural network to obtain the life prediction models of the wind power institutions in the multiple regions.

6. The method according to claim 5, characterized in that The method comprises: Verifying the life prediction models of wind power units in the multiple regions respectively to obtain multiple model prediction feedback data; Perform loss analysis based on the multiple model prediction feedback data to obtain multiple predicted loss data; The life prediction models of the wind power institutions in the plurality of regions are optimized based on the plurality of predicted loss data by using a model optimization algorithm to generate a plurality of optimized life prediction models of the wind power institutions in the plurality of regions.

7. A method for predicting the life of a wind power unit, characterized in that: include: Collecting target wind turbine mechanism operating data information of the target wind turbine and obtaining a global wind turbine mechanism life prediction model, wherein the global wind turbine mechanism life prediction model is trained by the wind turbine mechanism life prediction model training method according to any one of claims 1 to 6; The target wind power mechanism operation data information is predicted based on the global wind power mechanism life prediction model, a wind power mechanism life prediction result is output, and the wind power mechanism is maintained and managed based on the wind power mechanism life prediction result.

8. The method according to claim 7, characterized in that The maintenance and management of the wind power mechanism based on the life prediction result of the wind power mechanism includes: Build a wind power organization operation and maintenance solution library; Performing maintenance level analysis on the life prediction result of the wind power mechanism to determine maintenance level information of the wind power mechanism; Based on the matching of the wind power mechanism maintenance level information with the wind power mechanism operation and maintenance solution library, a target wind power mechanism operation and maintenance solution is obtained, and the wind power mechanism is maintained and managed according to the target wind power mechanism operation and maintenance solution.

9. A wind power mechanism life prediction model training device, characterized in that: The device comprises: The acquisition module is used to collect and obtain operating data information of multiple wind power units in multiple regional environments; An extraction module is used to extract correlation factors from the operating data information of the plurality of wind power units respectively to obtain life-span correlation feature information of the plurality of wind power units; A first training module is configured to input the life-related characteristic information of the plurality of wind power units as local sample data into a deep convolutional neural network for training, thereby obtaining life prediction models for wind power units in multiple regions; The second training module is used to encrypt and transmit the model parameters of the wind power plant life prediction models of the multiple regions to the federal central platform for global training to obtain a global wind power plant life prediction model.

10. A wind power mechanism life prediction device, characterized in that: include: an acquisition module, configured to collect target wind turbine mechanism operating data information of a target wind turbine and obtain a global wind turbine mechanism life prediction model, wherein the global wind turbine mechanism life prediction model is obtained by training using the wind turbine mechanism life prediction model training method according to any one of claims 1 to 7; The prediction module is used to predict the operating data information of the target wind power mechanism based on the global wind power mechanism life prediction model, output the wind power mechanism life prediction result, and maintain and manage the wind power mechanism based on the wind power mechanism life prediction result.

Images